1. Field of the Invention
The present invention relates to novel genes located in two chromosomal regions within E. coli that are associated with virulence. These chromosomal regions are known as pathogenicity islands (PAIs).
2. Related Background Art
Escherichia coli (E. coli) is a normal inhabitant of the intestine of humans and various animals. Pathogenic E. coli strains are able to cause infections of the intestine (intestinal E. coli strains) and of other organs such as the urinary tract (uropathogenic E. coli) or the brain (extraintestinal E coli). Intestinal pathogenic E. coli are a well established and leading cause of severe infantile diarrhea in the developing world. Additionally, cases of newborn meningitis and sepsis have been attributed to E. coli pathogens.
In contrast to non-pathogenic isolates, pathogenic E. coli produce pathogenicity factors which contribute to the ability of strains to cause infectious diseases (Muhldorfer, I. and Hacker, J., Microb. Pathogen. 16:171-181 1994). Adhesions facilitate binding of pathogenic bacteria to host tissues. Pathogenic E. coli strains also express toxins including haemolysins, which are involved in the destruction of host cells, and surface structures such as O-antigens, capsules or membrane proteins, which protect the bacteria from the action of phagocytes or the complement system (Ritter, et al., Mol. Microbiol. 17:109-212 1995).
The genes coding for pathogenicity factors of intestinal E. coli are located on large plasmids, phage genomes or on the chromosome. In contrast to intestinal E. coli, pathogenicity determinants of uropathogenic and other extraintestinal E. coli are, in most cases, located on the chromosome. Id.
Large chromosomal regions in pathogenic bacteria that encode adjacently located virulence genes have been termed pathogenicity islands ("PAIs"). PAIs are indicative of large fragments of DNA which comprise a group of virulence genes behaving as a distinct molecular and functional unit much like an island within the bacterial chromosome. For example, intact PAIs appear to transfer between organisms and confer complex virulence properties to the recipient bacteria.
Chromosomal PAIs in bacterial cells have been described in increasing detail over recent years. For example, J. Hacker and co-workers described two large, unstable regions in the chromosome of uropathogenic Escherichia coil strain 536 as PAI-I and PAI-II (Hacker J., et al., Microbiol. Pathog. 8:213-25 1990). Hacker found that PAI-I and PAI-II containing virulence regions can be lost by spontaneous deletion due to recombination events. Both of these PAIs were found to encode multiple virulence genes, and their loss resulted in reduced hemolytic activity, serum resistance, mannose-resistant hemagglutination, uroepithelial cell binding, and mouse virulence of the E. coli. (Knapp, S et al., J Bacteriol. 168:22-30 1986). Therefore, pathogenicity islands are characterized by their ability to confer complex virulence phenotypes to bacterial cells.
In addition to E. coli, specific deletion of large virulence regions has been observed in other bacteria such as Yersinia pestis. For example, Fetherston and co-workers found that a 102-kb region of the Y. pestis chromosome lost by spontaneous deletion resulted in the loss of many Y. pestis virulence phenotypes. (Fetherston, J. D. and Perry, R. D., MoL Microbiol. 13:697-708 1994, Fetherston, et al., Mol. Microbiol. 6:2693-704 1992). In this instance, the deletion appeared to be due to recombination within 2.2-kb repetitive elements at both ends of the 102-kb region.
It is possible that deletion of PAIs may benefit the organism by modulating bacterial virulence or genome size during infection. PAIs may also represent foreign DNA segments that were acquired during bacterial evolution that conferred important pathogenic properties to the bacteria. Observed flanking repeats, as observed in Y. pestis for example, may suggest a common mechanism by which these virulence genes were integrated into the bacterial chromosomes.
Integration of the virulence genes into bacterial chromosomes was further elucidated by the discovery and characterization of a locus of enterocyte effacement (the LEE locus) in enteropathogenic E. coli (McDaniel, et al., Proc. Natl Acad. Sci. (USA) 92:1664-8 1995). The LEE locus comprises 35-kb and encodes many genes required for these bacteria to "invade" and degrade the apical structure of enerocytes causing diarrhea. Although the LEE and PAI-I loci encode different virulence genes, these elements are located at the exact same site in the E. coli genome and contain the same DNA sequence within their right-hand ends, thus suggesting a common mechanism for their insertion.
Besides being found in enteropathogenic E. coli, the LEE element is also present in rabbit diarrheal E. coli, Hafnia alvei, and Citrobacter freundii biotype 4280, all of which induce attaching and effacing lesions on the apical face of enterocytes. The LEE locus appears to be inserted in the bacterial chromosome as a discrete molecular and functional virulence unit in the same fashion as PAI, PAI-II, and Yersinia PAI.
Along these same lines, a 40-kb Salmonella typhimurium PAI was characterized on the bacterial chromosome which encodes genes required for Salmonella entry into nonphagocytic epithelial cells of the intestine (Mills, D. M., et al., Mol. Microbiol. 15:749-59 1995). Like the LEE element, this PAI confers to Salmonella the ability to invade intestinal cells, and hence may likewise be characterized as an "invasion" PAI.
The pathogenicity islands described above all possess the common feature of conferring complex virulence properties to the recipient bacteria. However, they may be separated into two types by their respective contributions to virulence. PAI-I, PAI-II, and the Y. pestis PAI confer multiple virulence phenotypes, while the LEE and the S. typhimurium "invasion" PAI encode many genes specifying a single, complex virulence process.
It is advantageous to characterize closely-related bacteria that contain or do not contain the PAI by the isolation of a discrete molecular and functional unit on the bacterial chromosome. Since the presence versus the absence of essential virulence genes can often distinguish closely-related virulent versus avirulent bacterial strains or species, experiments have been conducted to identify virulence loci and potential PAIs by isolating DNA sequences that are unique to virulent bacteria (Bloch, C. A., et al, J Bacteriol. 176:7121-5 1994, Groisman, E. A., EMBO J. 12:3779-87 1993).
At least two PAIs are present in E. coli J96. These PAIs, PAI:IV and PAI V are linked to tRNA loci but at sites different from those occupied by other known E. coli PAIs. Swenson et al., Infect. and Immun. 64:3736-3743 (1996).
The era of true comparative genomics has been ushered in by high through-put genomic sequencing and analysis. The first two complete bacterial genome sequences, those of Haemophilus influenzae and Mycoplasma genitalium were recently described (Fleischmann, R. D., et al., Science 269:496 (1995); Fraser, C. M., et al., Science 270:397 (1995)). Large scale DNA sequencing efforts also have produced an extensive collection of sequence data from eukaryotes, including Homo sapiens (Adams, M. D., et al., Nature 377:3 (1995)) and Saccharomyces cerevisiae (Levy, J., Yeast 10:1689 (1994)).
The need continues to exist for the application of high through-put sequencing and analysis to study genomes and subgenomes of infectious organisms. Further, a need exists for genetic markers that can be employed to distinguish closely-related virulent and avirulent strains of a given bacteria.